Here is the problem.
| # | User | Rating |
|---|---|---|
| 1 | Benq | 3676 |
| 2 | Kevin114514 | 3659 |
| 3 | jiangly | 3640 |
| 4 | turmax | 3450 |
| 5 | tourist | 3382 |
| 6 | strapple | 3360 |
| 7 | Um_nik | 3354 |
| 8 | squareOf105 | 3348 |
| 9 | maroonrk | 3312 |
| 10 | idontwannadothis…e | 3191 |
| # | User | Contrib. |
|---|---|---|
| 1 | Qingyu | 156 |
| 2 | nik_exists | 151 |
| 3 | maspy | 149 |
| 4 | AmShZ | 145 |
| 5 | Um_nik | 143 |
| 6 | Errichto | 139 |
| 7 | BledDest | 136 |
| 8 | adamant | 135 |
| 9 | maroonrk | 132 |
| 10 | cry | 130 |
Here is the problem.
| Name |
|---|



Hi! Let f[i={1,...,N}][j={1,...,M}] be the number of ways if you have the first i children and the i-th has j candies. That is M*N=10^12 states in total. My approach uses a state f[i={1,...,N}][j={1,...sqrt(M)}][k={0,1}] that is the answer if we have the first i children and if k=0 j is the number of the candies that were given to the i-th child. If k=1 j is the maximum number of candies that you can give to the (i-1)-th child, so if you are in state (i,j,1) you know that you will give M/j to the i-th child.
f[1][1,...,sqrt(M)][0,1]=1
f[i][j][0]=number of ways (i-1)th child can take 1,...,M/j candies.
f[i][j][1]=number of ways (i-1)th child can take 1,...,j candies.
So:
f[i][j][0]=f[i-1][j][0]+...+f[i-1][min(sqrt(M),M/j)][0]+f[i-1][1][1]+...+f[i-1][j][1]
and
f[i][j][1]=f[i-1][j][0]+...+f[i-1][min(sqrt(M),j)][0]+f[i-1][1][1]+...+f[i-1][M/j][1]
You can compute each state in O(1) if you have the sums from the last step. I hope I didn't make a mistake, but if I did you have the idea!